SD-squared revisited: reply to Coltheart, Tree, and Saunders (2010).